razhan commited on
Commit
1eeb8dd
1 Parent(s): 7717abe

End of training

Browse files
Files changed (4) hide show
  1. all_results.json +12 -12
  2. eval_results.json +7 -7
  3. train_results.json +6 -6
  4. trainer_state.json +1676 -14
all_results.json CHANGED
@@ -1,19 +1,19 @@
1
  {
2
- "epoch": 3.0,
3
- "eval_cer": 10.7576,
4
- "eval_gen_len": 16.9735,
5
- "eval_loss": 4.2736406326293945,
6
  "eval_rouge1": 0.0,
7
  "eval_rouge2": 0.0,
8
  "eval_rougeL": 0.0,
9
  "eval_rougeLsum": 0.0,
10
- "eval_runtime": 37.6133,
11
  "eval_samples": 2000,
12
- "eval_samples_per_second": 53.173,
13
- "eval_steps_per_second": 0.425,
14
- "train_loss": 4.75333997938368,
15
- "train_runtime": 38.4881,
16
- "train_samples": 10000,
17
- "train_samples_per_second": 779.461,
18
- "train_steps_per_second": 3.508
19
  }
 
1
  {
2
+ "epoch": 5.0,
3
+ "eval_cer": 4.7349,
4
+ "eval_gen_len": 13.2035,
5
+ "eval_loss": 0.6352977752685547,
6
  "eval_rouge1": 0.0,
7
  "eval_rouge2": 0.0,
8
  "eval_rougeL": 0.0,
9
  "eval_rougeLsum": 0.0,
10
+ "eval_runtime": 27.1836,
11
  "eval_samples": 2000,
12
+ "eval_samples_per_second": 73.574,
13
+ "eval_steps_per_second": 0.294,
14
+ "train_loss": 0.564045099207411,
15
+ "train_runtime": 5933.1558,
16
+ "train_samples": 1471386,
17
+ "train_samples_per_second": 1239.969,
18
+ "train_steps_per_second": 3.876
19
  }
eval_results.json CHANGED
@@ -1,14 +1,14 @@
1
  {
2
- "epoch": 3.0,
3
- "eval_cer": 10.7576,
4
- "eval_gen_len": 16.9735,
5
- "eval_loss": 4.2736406326293945,
6
  "eval_rouge1": 0.0,
7
  "eval_rouge2": 0.0,
8
  "eval_rougeL": 0.0,
9
  "eval_rougeLsum": 0.0,
10
- "eval_runtime": 37.6133,
11
  "eval_samples": 2000,
12
- "eval_samples_per_second": 53.173,
13
- "eval_steps_per_second": 0.425
14
  }
 
1
  {
2
+ "epoch": 5.0,
3
+ "eval_cer": 4.7349,
4
+ "eval_gen_len": 13.2035,
5
+ "eval_loss": 0.6352977752685547,
6
  "eval_rouge1": 0.0,
7
  "eval_rouge2": 0.0,
8
  "eval_rougeL": 0.0,
9
  "eval_rougeLsum": 0.0,
10
+ "eval_runtime": 27.1836,
11
  "eval_samples": 2000,
12
+ "eval_samples_per_second": 73.574,
13
+ "eval_steps_per_second": 0.294
14
  }
train_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
- "epoch": 3.0,
3
- "train_loss": 4.75333997938368,
4
- "train_runtime": 38.4881,
5
- "train_samples": 10000,
6
- "train_samples_per_second": 779.461,
7
- "train_steps_per_second": 3.508
8
  }
 
1
  {
2
+ "epoch": 5.0,
3
+ "train_loss": 0.564045099207411,
4
+ "train_runtime": 5933.1558,
5
+ "train_samples": 1471386,
6
+ "train_samples_per_second": 1239.969,
7
+ "train_steps_per_second": 3.876
8
  }
trainer_state.json CHANGED
@@ -1,34 +1,1696 @@
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
- "epoch": 3.0,
5
  "eval_steps": 500,
6
- "global_step": 135,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 2.22,
13
- "learning_rate": 1.4074074074074075e-05,
14
- "loss": 4.9042,
15
  "step": 100
16
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
17
  {
18
  "epoch": 3.0,
19
- "step": 135,
20
- "total_flos": 719870106009600.0,
21
- "train_loss": 4.75333997938368,
22
- "train_runtime": 38.4881,
23
- "train_samples_per_second": 779.461,
24
- "train_steps_per_second": 3.508
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
25
  }
26
  ],
27
  "logging_steps": 100,
28
- "max_steps": 135,
29
- "num_train_epochs": 3,
30
  "save_steps": 500,
31
- "total_flos": 719870106009600.0,
32
  "trial_name": null,
33
  "trial_params": null
34
  }
 
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
+ "epoch": 5.0,
5
  "eval_steps": 500,
6
+ "global_step": 22995,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.02,
13
+ "learning_rate": 4.9648474306008555e-05,
14
+ "loss": 4.8007,
15
  "step": 100
16
  },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 4.9286076683336956e-05,
20
+ "loss": 4.0366,
21
+ "step": 200
22
+ },
23
+ {
24
+ "epoch": 0.07,
25
+ "learning_rate": 4.892367906066536e-05,
26
+ "loss": 3.8316,
27
+ "step": 300
28
+ },
29
+ {
30
+ "epoch": 0.09,
31
+ "learning_rate": 4.8561281437993764e-05,
32
+ "loss": 3.689,
33
+ "step": 400
34
+ },
35
+ {
36
+ "epoch": 0.11,
37
+ "learning_rate": 4.819888381532217e-05,
38
+ "loss": 3.5765,
39
+ "step": 500
40
+ },
41
+ {
42
+ "epoch": 0.11,
43
+ "eval_cer": 10.3648,
44
+ "eval_gen_len": 16.358,
45
+ "eval_loss": 3.433837413787842,
46
+ "eval_rouge1": 0.0,
47
+ "eval_rouge2": 0.0,
48
+ "eval_rougeL": 0.0,
49
+ "eval_rougeLsum": 0.0,
50
+ "eval_runtime": 38.4477,
51
+ "eval_samples_per_second": 52.019,
52
+ "eval_steps_per_second": 0.208,
53
+ "step": 500
54
+ },
55
+ {
56
+ "epoch": 0.13,
57
+ "learning_rate": 4.783648619265057e-05,
58
+ "loss": 3.484,
59
+ "step": 600
60
+ },
61
+ {
62
+ "epoch": 0.15,
63
+ "learning_rate": 4.747408856997898e-05,
64
+ "loss": 3.3864,
65
+ "step": 700
66
+ },
67
+ {
68
+ "epoch": 0.17,
69
+ "learning_rate": 4.711169094730739e-05,
70
+ "loss": 3.3142,
71
+ "step": 800
72
+ },
73
+ {
74
+ "epoch": 0.2,
75
+ "learning_rate": 4.674929332463579e-05,
76
+ "loss": 3.2487,
77
+ "step": 900
78
+ },
79
+ {
80
+ "epoch": 0.22,
81
+ "learning_rate": 4.63868957019642e-05,
82
+ "loss": 3.1749,
83
+ "step": 1000
84
+ },
85
+ {
86
+ "epoch": 0.22,
87
+ "eval_cer": 10.955,
88
+ "eval_gen_len": 15.828,
89
+ "eval_loss": 3.0109550952911377,
90
+ "eval_rouge1": 0.0,
91
+ "eval_rouge2": 0.0,
92
+ "eval_rougeL": 0.0,
93
+ "eval_rougeLsum": 0.0,
94
+ "eval_runtime": 38.0142,
95
+ "eval_samples_per_second": 52.612,
96
+ "eval_steps_per_second": 0.21,
97
+ "step": 1000
98
+ },
99
+ {
100
+ "epoch": 0.24,
101
+ "learning_rate": 4.60244980792926e-05,
102
+ "loss": 3.1153,
103
+ "step": 1100
104
+ },
105
+ {
106
+ "epoch": 0.26,
107
+ "learning_rate": 4.5662100456621006e-05,
108
+ "loss": 3.0549,
109
+ "step": 1200
110
+ },
111
+ {
112
+ "epoch": 0.28,
113
+ "learning_rate": 4.529970283394941e-05,
114
+ "loss": 2.9852,
115
+ "step": 1300
116
+ },
117
+ {
118
+ "epoch": 0.3,
119
+ "learning_rate": 4.4937305211277815e-05,
120
+ "loss": 2.9315,
121
+ "step": 1400
122
+ },
123
+ {
124
+ "epoch": 0.33,
125
+ "learning_rate": 4.4574907588606215e-05,
126
+ "loss": 2.8696,
127
+ "step": 1500
128
+ },
129
+ {
130
+ "epoch": 0.33,
131
+ "eval_cer": 11.2579,
132
+ "eval_gen_len": 15.358,
133
+ "eval_loss": 2.6943440437316895,
134
+ "eval_rouge1": 0.0,
135
+ "eval_rouge2": 0.0,
136
+ "eval_rougeL": 0.0,
137
+ "eval_rougeLsum": 0.0,
138
+ "eval_runtime": 36.6296,
139
+ "eval_samples_per_second": 54.601,
140
+ "eval_steps_per_second": 0.218,
141
+ "step": 1500
142
+ },
143
+ {
144
+ "epoch": 0.35,
145
+ "learning_rate": 4.421250996593462e-05,
146
+ "loss": 2.8218,
147
+ "step": 1600
148
+ },
149
+ {
150
+ "epoch": 0.37,
151
+ "learning_rate": 4.385011234326303e-05,
152
+ "loss": 2.7759,
153
+ "step": 1700
154
+ },
155
+ {
156
+ "epoch": 0.39,
157
+ "learning_rate": 4.348771472059143e-05,
158
+ "loss": 2.7254,
159
+ "step": 1800
160
+ },
161
+ {
162
+ "epoch": 0.41,
163
+ "learning_rate": 4.312531709791984e-05,
164
+ "loss": 2.6728,
165
+ "step": 1900
166
+ },
167
+ {
168
+ "epoch": 0.43,
169
+ "learning_rate": 4.276291947524824e-05,
170
+ "loss": 2.6279,
171
+ "step": 2000
172
+ },
173
+ {
174
+ "epoch": 0.43,
175
+ "eval_cer": 11.6734,
176
+ "eval_gen_len": 14.8565,
177
+ "eval_loss": 2.4299662113189697,
178
+ "eval_rouge1": 0.0,
179
+ "eval_rouge2": 0.0,
180
+ "eval_rougeL": 0.0,
181
+ "eval_rougeLsum": 0.0,
182
+ "eval_runtime": 35.0111,
183
+ "eval_samples_per_second": 57.125,
184
+ "eval_steps_per_second": 0.228,
185
+ "step": 2000
186
+ },
187
+ {
188
+ "epoch": 0.46,
189
+ "learning_rate": 4.240052185257665e-05,
190
+ "loss": 2.5848,
191
+ "step": 2100
192
+ },
193
+ {
194
+ "epoch": 0.48,
195
+ "learning_rate": 4.203812422990505e-05,
196
+ "loss": 2.5407,
197
+ "step": 2200
198
+ },
199
+ {
200
+ "epoch": 0.5,
201
+ "learning_rate": 4.167572660723346e-05,
202
+ "loss": 2.4987,
203
+ "step": 2300
204
+ },
205
+ {
206
+ "epoch": 0.52,
207
+ "learning_rate": 4.131332898456186e-05,
208
+ "loss": 2.4587,
209
+ "step": 2400
210
+ },
211
+ {
212
+ "epoch": 0.54,
213
+ "learning_rate": 4.0950931361890266e-05,
214
+ "loss": 2.4166,
215
+ "step": 2500
216
+ },
217
+ {
218
+ "epoch": 0.54,
219
+ "eval_cer": 11.028,
220
+ "eval_gen_len": 14.523,
221
+ "eval_loss": 2.224823474884033,
222
+ "eval_rouge1": 0.0,
223
+ "eval_rouge2": 0.0,
224
+ "eval_rougeL": 0.0,
225
+ "eval_rougeLsum": 0.0,
226
+ "eval_runtime": 32.1586,
227
+ "eval_samples_per_second": 62.192,
228
+ "eval_steps_per_second": 0.249,
229
+ "step": 2500
230
+ },
231
+ {
232
+ "epoch": 0.57,
233
+ "learning_rate": 4.058853373921867e-05,
234
+ "loss": 2.3746,
235
+ "step": 2600
236
+ },
237
+ {
238
+ "epoch": 0.59,
239
+ "learning_rate": 4.0226136116547074e-05,
240
+ "loss": 2.3435,
241
+ "step": 2700
242
+ },
243
+ {
244
+ "epoch": 0.61,
245
+ "learning_rate": 3.986373849387548e-05,
246
+ "loss": 2.3134,
247
+ "step": 2800
248
+ },
249
+ {
250
+ "epoch": 0.63,
251
+ "learning_rate": 3.950134087120388e-05,
252
+ "loss": 2.2865,
253
+ "step": 2900
254
+ },
255
+ {
256
+ "epoch": 0.65,
257
+ "learning_rate": 3.913894324853229e-05,
258
+ "loss": 2.2444,
259
+ "step": 3000
260
+ },
261
+ {
262
+ "epoch": 0.65,
263
+ "eval_cer": 10.4379,
264
+ "eval_gen_len": 14.4185,
265
+ "eval_loss": 2.0489799976348877,
266
+ "eval_rouge1": 0.0,
267
+ "eval_rouge2": 0.0,
268
+ "eval_rougeL": 0.0,
269
+ "eval_rougeLsum": 0.0,
270
+ "eval_runtime": 32.7499,
271
+ "eval_samples_per_second": 61.069,
272
+ "eval_steps_per_second": 0.244,
273
+ "step": 3000
274
+ },
275
+ {
276
+ "epoch": 0.67,
277
+ "learning_rate": 3.877654562586069e-05,
278
+ "loss": 2.2154,
279
+ "step": 3100
280
+ },
281
+ {
282
+ "epoch": 0.7,
283
+ "learning_rate": 3.84141480031891e-05,
284
+ "loss": 2.1757,
285
+ "step": 3200
286
+ },
287
+ {
288
+ "epoch": 0.72,
289
+ "learning_rate": 3.80517503805175e-05,
290
+ "loss": 2.1584,
291
+ "step": 3300
292
+ },
293
+ {
294
+ "epoch": 0.74,
295
+ "learning_rate": 3.768935275784591e-05,
296
+ "loss": 2.1166,
297
+ "step": 3400
298
+ },
299
+ {
300
+ "epoch": 0.76,
301
+ "learning_rate": 3.7326955135174316e-05,
302
+ "loss": 2.0902,
303
+ "step": 3500
304
+ },
305
+ {
306
+ "epoch": 0.76,
307
+ "eval_cer": 10.0905,
308
+ "eval_gen_len": 14.242,
309
+ "eval_loss": 1.8756157159805298,
310
+ "eval_rouge1": 0.0,
311
+ "eval_rouge2": 0.0,
312
+ "eval_rougeL": 0.0,
313
+ "eval_rougeLsum": 0.0,
314
+ "eval_runtime": 32.3794,
315
+ "eval_samples_per_second": 61.768,
316
+ "eval_steps_per_second": 0.247,
317
+ "step": 3500
318
+ },
319
+ {
320
+ "epoch": 0.78,
321
+ "learning_rate": 3.696455751250272e-05,
322
+ "loss": 2.057,
323
+ "step": 3600
324
+ },
325
+ {
326
+ "epoch": 0.8,
327
+ "learning_rate": 3.6602159889831124e-05,
328
+ "loss": 2.0307,
329
+ "step": 3700
330
+ },
331
+ {
332
+ "epoch": 0.83,
333
+ "learning_rate": 3.6243386243386245e-05,
334
+ "loss": 2.0145,
335
+ "step": 3800
336
+ },
337
+ {
338
+ "epoch": 0.85,
339
+ "learning_rate": 3.588098862071465e-05,
340
+ "loss": 1.9898,
341
+ "step": 3900
342
+ },
343
+ {
344
+ "epoch": 0.87,
345
+ "learning_rate": 3.5518590998043053e-05,
346
+ "loss": 1.9565,
347
+ "step": 4000
348
+ },
349
+ {
350
+ "epoch": 0.87,
351
+ "eval_cer": 9.7629,
352
+ "eval_gen_len": 14.042,
353
+ "eval_loss": 1.7376880645751953,
354
+ "eval_rouge1": 0.0,
355
+ "eval_rouge2": 0.0,
356
+ "eval_rougeL": 0.0,
357
+ "eval_rougeLsum": 0.0,
358
+ "eval_runtime": 31.0968,
359
+ "eval_samples_per_second": 64.315,
360
+ "eval_steps_per_second": 0.257,
361
+ "step": 4000
362
+ },
363
+ {
364
+ "epoch": 0.89,
365
+ "learning_rate": 3.515619337537146e-05,
366
+ "loss": 1.921,
367
+ "step": 4100
368
+ },
369
+ {
370
+ "epoch": 0.91,
371
+ "learning_rate": 3.479379575269986e-05,
372
+ "loss": 1.9099,
373
+ "step": 4200
374
+ },
375
+ {
376
+ "epoch": 0.93,
377
+ "learning_rate": 3.443139813002827e-05,
378
+ "loss": 1.8779,
379
+ "step": 4300
380
+ },
381
+ {
382
+ "epoch": 0.96,
383
+ "learning_rate": 3.406900050735668e-05,
384
+ "loss": 1.8613,
385
+ "step": 4400
386
+ },
387
+ {
388
+ "epoch": 0.98,
389
+ "learning_rate": 3.370660288468508e-05,
390
+ "loss": 1.8319,
391
+ "step": 4500
392
+ },
393
+ {
394
+ "epoch": 0.98,
395
+ "eval_cer": 9.4737,
396
+ "eval_gen_len": 13.877,
397
+ "eval_loss": 1.6244360208511353,
398
+ "eval_rouge1": 0.0,
399
+ "eval_rouge2": 0.0,
400
+ "eval_rougeL": 0.0,
401
+ "eval_rougeLsum": 0.0,
402
+ "eval_runtime": 31.2035,
403
+ "eval_samples_per_second": 64.095,
404
+ "eval_steps_per_second": 0.256,
405
+ "step": 4500
406
+ },
407
+ {
408
+ "epoch": 1.0,
409
+ "learning_rate": 3.3344205262013486e-05,
410
+ "loss": 1.8114,
411
+ "step": 4600
412
+ },
413
+ {
414
+ "epoch": 1.02,
415
+ "learning_rate": 3.298180763934189e-05,
416
+ "loss": 1.7689,
417
+ "step": 4700
418
+ },
419
+ {
420
+ "epoch": 1.04,
421
+ "learning_rate": 3.2619410016670295e-05,
422
+ "loss": 1.7452,
423
+ "step": 4800
424
+ },
425
+ {
426
+ "epoch": 1.07,
427
+ "learning_rate": 3.2257012393998696e-05,
428
+ "loss": 1.7304,
429
+ "step": 4900
430
+ },
431
+ {
432
+ "epoch": 1.09,
433
+ "learning_rate": 3.1894614771327104e-05,
434
+ "loss": 1.7104,
435
+ "step": 5000
436
+ },
437
+ {
438
+ "epoch": 1.09,
439
+ "eval_cer": 9.153,
440
+ "eval_gen_len": 13.825,
441
+ "eval_loss": 1.5353318452835083,
442
+ "eval_rouge1": 0.0,
443
+ "eval_rouge2": 0.0,
444
+ "eval_rougeL": 0.0,
445
+ "eval_rougeLsum": 0.0,
446
+ "eval_runtime": 30.755,
447
+ "eval_samples_per_second": 65.03,
448
+ "eval_steps_per_second": 0.26,
449
+ "step": 5000
450
+ },
451
+ {
452
+ "epoch": 1.11,
453
+ "learning_rate": 3.1532217148655505e-05,
454
+ "loss": 1.6878,
455
+ "step": 5100
456
+ },
457
+ {
458
+ "epoch": 1.13,
459
+ "learning_rate": 3.116981952598391e-05,
460
+ "loss": 1.6728,
461
+ "step": 5200
462
+ },
463
+ {
464
+ "epoch": 1.15,
465
+ "learning_rate": 3.080742190331232e-05,
466
+ "loss": 1.6593,
467
+ "step": 5300
468
+ },
469
+ {
470
+ "epoch": 1.17,
471
+ "learning_rate": 3.044502428064072e-05,
472
+ "loss": 1.6418,
473
+ "step": 5400
474
+ },
475
+ {
476
+ "epoch": 1.2,
477
+ "learning_rate": 3.0082626657969125e-05,
478
+ "loss": 1.6148,
479
+ "step": 5500
480
+ },
481
+ {
482
+ "epoch": 1.2,
483
+ "eval_cer": 8.8125,
484
+ "eval_gen_len": 13.726,
485
+ "eval_loss": 1.4422341585159302,
486
+ "eval_rouge1": 0.0,
487
+ "eval_rouge2": 0.0,
488
+ "eval_rougeL": 0.0,
489
+ "eval_rougeLsum": 0.0,
490
+ "eval_runtime": 30.1297,
491
+ "eval_samples_per_second": 66.38,
492
+ "eval_steps_per_second": 0.266,
493
+ "step": 5500
494
+ },
495
+ {
496
+ "epoch": 1.22,
497
+ "learning_rate": 2.972022903529753e-05,
498
+ "loss": 1.6163,
499
+ "step": 5600
500
+ },
501
+ {
502
+ "epoch": 1.24,
503
+ "learning_rate": 2.9357831412625937e-05,
504
+ "loss": 1.5967,
505
+ "step": 5700
506
+ },
507
+ {
508
+ "epoch": 1.26,
509
+ "learning_rate": 2.8995433789954342e-05,
510
+ "loss": 1.5853,
511
+ "step": 5800
512
+ },
513
+ {
514
+ "epoch": 1.28,
515
+ "learning_rate": 2.8633036167282746e-05,
516
+ "loss": 1.5532,
517
+ "step": 5900
518
+ },
519
+ {
520
+ "epoch": 1.3,
521
+ "learning_rate": 2.827063854461115e-05,
522
+ "loss": 1.5459,
523
+ "step": 6000
524
+ },
525
+ {
526
+ "epoch": 1.3,
527
+ "eval_cer": 8.5589,
528
+ "eval_gen_len": 13.681,
529
+ "eval_loss": 1.3754663467407227,
530
+ "eval_rouge1": 0.0,
531
+ "eval_rouge2": 0.0,
532
+ "eval_rougeL": 0.0,
533
+ "eval_rougeLsum": 0.0,
534
+ "eval_runtime": 30.0363,
535
+ "eval_samples_per_second": 66.586,
536
+ "eval_steps_per_second": 0.266,
537
+ "step": 6000
538
+ },
539
+ {
540
+ "epoch": 1.33,
541
+ "learning_rate": 2.7908240921939555e-05,
542
+ "loss": 1.5384,
543
+ "step": 6100
544
+ },
545
+ {
546
+ "epoch": 1.35,
547
+ "learning_rate": 2.754584329926796e-05,
548
+ "loss": 1.5169,
549
+ "step": 6200
550
+ },
551
+ {
552
+ "epoch": 1.37,
553
+ "learning_rate": 2.7183445676596363e-05,
554
+ "loss": 1.5045,
555
+ "step": 6300
556
+ },
557
+ {
558
+ "epoch": 1.39,
559
+ "learning_rate": 2.6821048053924768e-05,
560
+ "loss": 1.4945,
561
+ "step": 6400
562
+ },
563
+ {
564
+ "epoch": 1.41,
565
+ "learning_rate": 2.6458650431253172e-05,
566
+ "loss": 1.4837,
567
+ "step": 6500
568
+ },
569
+ {
570
+ "epoch": 1.41,
571
+ "eval_cer": 8.2717,
572
+ "eval_gen_len": 13.6225,
573
+ "eval_loss": 1.3035191297531128,
574
+ "eval_rouge1": 0.0,
575
+ "eval_rouge2": 0.0,
576
+ "eval_rougeL": 0.0,
577
+ "eval_rougeLsum": 0.0,
578
+ "eval_runtime": 29.7288,
579
+ "eval_samples_per_second": 67.275,
580
+ "eval_steps_per_second": 0.269,
581
+ "step": 6500
582
+ },
583
+ {
584
+ "epoch": 1.44,
585
+ "learning_rate": 2.609625280858158e-05,
586
+ "loss": 1.4608,
587
+ "step": 6600
588
+ },
589
+ {
590
+ "epoch": 1.46,
591
+ "learning_rate": 2.5733855185909984e-05,
592
+ "loss": 1.4631,
593
+ "step": 6700
594
+ },
595
+ {
596
+ "epoch": 1.48,
597
+ "learning_rate": 2.537145756323839e-05,
598
+ "loss": 1.454,
599
+ "step": 6800
600
+ },
601
+ {
602
+ "epoch": 1.5,
603
+ "learning_rate": 2.5009059940566793e-05,
604
+ "loss": 1.4402,
605
+ "step": 6900
606
+ },
607
+ {
608
+ "epoch": 1.52,
609
+ "learning_rate": 2.4646662317895194e-05,
610
+ "loss": 1.4219,
611
+ "step": 7000
612
+ },
613
+ {
614
+ "epoch": 1.52,
615
+ "eval_cer": 8.0684,
616
+ "eval_gen_len": 13.549,
617
+ "eval_loss": 1.2407491207122803,
618
+ "eval_rouge1": 0.0,
619
+ "eval_rouge2": 0.0,
620
+ "eval_rougeL": 0.0,
621
+ "eval_rougeLsum": 0.0,
622
+ "eval_runtime": 29.9599,
623
+ "eval_samples_per_second": 66.756,
624
+ "eval_steps_per_second": 0.267,
625
+ "step": 7000
626
+ },
627
+ {
628
+ "epoch": 1.54,
629
+ "learning_rate": 2.4284264695223598e-05,
630
+ "loss": 1.4081,
631
+ "step": 7100
632
+ },
633
+ {
634
+ "epoch": 1.57,
635
+ "learning_rate": 2.3921867072552002e-05,
636
+ "loss": 1.3969,
637
+ "step": 7200
638
+ },
639
+ {
640
+ "epoch": 1.59,
641
+ "learning_rate": 2.355946944988041e-05,
642
+ "loss": 1.3942,
643
+ "step": 7300
644
+ },
645
+ {
646
+ "epoch": 1.61,
647
+ "learning_rate": 2.3197071827208815e-05,
648
+ "loss": 1.3834,
649
+ "step": 7400
650
+ },
651
+ {
652
+ "epoch": 1.63,
653
+ "learning_rate": 2.283467420453722e-05,
654
+ "loss": 1.3743,
655
+ "step": 7500
656
+ },
657
+ {
658
+ "epoch": 1.63,
659
+ "eval_cer": 7.7684,
660
+ "eval_gen_len": 13.502,
661
+ "eval_loss": 1.186456561088562,
662
+ "eval_rouge1": 0.0,
663
+ "eval_rouge2": 0.0,
664
+ "eval_rougeL": 0.0,
665
+ "eval_rougeLsum": 0.0,
666
+ "eval_runtime": 29.5504,
667
+ "eval_samples_per_second": 67.681,
668
+ "eval_steps_per_second": 0.271,
669
+ "step": 7500
670
+ },
671
+ {
672
+ "epoch": 1.65,
673
+ "learning_rate": 2.2472276581865623e-05,
674
+ "loss": 1.361,
675
+ "step": 7600
676
+ },
677
+ {
678
+ "epoch": 1.67,
679
+ "learning_rate": 2.2109878959194028e-05,
680
+ "loss": 1.3531,
681
+ "step": 7700
682
+ },
683
+ {
684
+ "epoch": 1.7,
685
+ "learning_rate": 2.1747481336522432e-05,
686
+ "loss": 1.3436,
687
+ "step": 7800
688
+ },
689
+ {
690
+ "epoch": 1.72,
691
+ "learning_rate": 2.1385083713850836e-05,
692
+ "loss": 1.3365,
693
+ "step": 7900
694
+ },
695
+ {
696
+ "epoch": 1.74,
697
+ "learning_rate": 2.102268609117924e-05,
698
+ "loss": 1.3318,
699
+ "step": 8000
700
+ },
701
+ {
702
+ "epoch": 1.74,
703
+ "eval_cer": 7.5247,
704
+ "eval_gen_len": 13.509,
705
+ "eval_loss": 1.1502524614334106,
706
+ "eval_rouge1": 0.0,
707
+ "eval_rouge2": 0.0,
708
+ "eval_rougeL": 0.0,
709
+ "eval_rougeLsum": 0.0,
710
+ "eval_runtime": 29.2293,
711
+ "eval_samples_per_second": 68.424,
712
+ "eval_steps_per_second": 0.274,
713
+ "step": 8000
714
+ },
715
+ {
716
+ "epoch": 1.76,
717
+ "learning_rate": 2.0660288468507645e-05,
718
+ "loss": 1.3248,
719
+ "step": 8100
720
+ },
721
+ {
722
+ "epoch": 1.78,
723
+ "learning_rate": 2.0297890845836053e-05,
724
+ "loss": 1.312,
725
+ "step": 8200
726
+ },
727
+ {
728
+ "epoch": 1.8,
729
+ "learning_rate": 1.9935493223164457e-05,
730
+ "loss": 1.3043,
731
+ "step": 8300
732
+ },
733
+ {
734
+ "epoch": 1.83,
735
+ "learning_rate": 1.957309560049286e-05,
736
+ "loss": 1.298,
737
+ "step": 8400
738
+ },
739
+ {
740
+ "epoch": 1.85,
741
+ "learning_rate": 1.9210697977821266e-05,
742
+ "loss": 1.2893,
743
+ "step": 8500
744
+ },
745
+ {
746
+ "epoch": 1.85,
747
+ "eval_cer": 7.3826,
748
+ "eval_gen_len": 13.456,
749
+ "eval_loss": 1.108539342880249,
750
+ "eval_rouge1": 0.0,
751
+ "eval_rouge2": 0.0,
752
+ "eval_rougeL": 0.0,
753
+ "eval_rougeLsum": 0.0,
754
+ "eval_runtime": 29.5564,
755
+ "eval_samples_per_second": 67.667,
756
+ "eval_steps_per_second": 0.271,
757
+ "step": 8500
758
+ },
759
+ {
760
+ "epoch": 1.87,
761
+ "learning_rate": 3.1308980213089806e-05,
762
+ "loss": 1.2603,
763
+ "step": 8600
764
+ },
765
+ {
766
+ "epoch": 1.89,
767
+ "learning_rate": 3.109154163948685e-05,
768
+ "loss": 1.2545,
769
+ "step": 8700
770
+ },
771
+ {
772
+ "epoch": 1.91,
773
+ "learning_rate": 3.0874103065883894e-05,
774
+ "loss": 1.2411,
775
+ "step": 8800
776
+ },
777
+ {
778
+ "epoch": 1.94,
779
+ "learning_rate": 3.0656664492280935e-05,
780
+ "loss": 1.2354,
781
+ "step": 8900
782
+ },
783
+ {
784
+ "epoch": 1.96,
785
+ "learning_rate": 3.043922591867798e-05,
786
+ "loss": 1.2243,
787
+ "step": 9000
788
+ },
789
+ {
790
+ "epoch": 1.98,
791
+ "learning_rate": 3.022178734507502e-05,
792
+ "loss": 1.2228,
793
+ "step": 9100
794
+ },
795
+ {
796
+ "epoch": 2.0,
797
+ "eval_cer": 7.0411,
798
+ "eval_gen_len": 13.3935,
799
+ "eval_loss": 1.0505541563034058,
800
+ "eval_rouge1": 0.0,
801
+ "eval_rouge2": 0.0,
802
+ "eval_rougeL": 0.0,
803
+ "eval_rougeLsum": 0.0,
804
+ "eval_runtime": 29.0334,
805
+ "eval_samples_per_second": 68.886,
806
+ "eval_steps_per_second": 0.276,
807
+ "step": 9198
808
+ },
809
+ {
810
+ "epoch": 2.0,
811
+ "learning_rate": 3.0004348771472063e-05,
812
+ "loss": 1.2108,
813
+ "step": 9200
814
+ },
815
+ {
816
+ "epoch": 2.02,
817
+ "learning_rate": 2.9786910197869107e-05,
818
+ "loss": 1.1899,
819
+ "step": 9300
820
+ },
821
+ {
822
+ "epoch": 2.04,
823
+ "learning_rate": 2.9569471624266148e-05,
824
+ "loss": 1.1901,
825
+ "step": 9400
826
+ },
827
+ {
828
+ "epoch": 2.07,
829
+ "learning_rate": 2.9352033050663192e-05,
830
+ "loss": 1.1772,
831
+ "step": 9500
832
+ },
833
+ {
834
+ "epoch": 2.09,
835
+ "learning_rate": 2.9134594477060236e-05,
836
+ "loss": 1.1657,
837
+ "step": 9600
838
+ },
839
+ {
840
+ "epoch": 2.11,
841
+ "learning_rate": 2.8917155903457276e-05,
842
+ "loss": 1.1618,
843
+ "step": 9700
844
+ },
845
+ {
846
+ "epoch": 2.13,
847
+ "learning_rate": 2.869971732985432e-05,
848
+ "loss": 1.1517,
849
+ "step": 9800
850
+ },
851
+ {
852
+ "epoch": 2.15,
853
+ "learning_rate": 2.8482278756251364e-05,
854
+ "loss": 1.1428,
855
+ "step": 9900
856
+ },
857
+ {
858
+ "epoch": 2.17,
859
+ "learning_rate": 2.826701456838443e-05,
860
+ "loss": 1.1386,
861
+ "step": 10000
862
+ },
863
+ {
864
+ "epoch": 2.2,
865
+ "learning_rate": 2.8049575994781475e-05,
866
+ "loss": 1.1285,
867
+ "step": 10100
868
+ },
869
+ {
870
+ "epoch": 2.22,
871
+ "learning_rate": 2.7832137421178516e-05,
872
+ "loss": 1.1213,
873
+ "step": 10200
874
+ },
875
+ {
876
+ "epoch": 2.24,
877
+ "learning_rate": 2.761469884757556e-05,
878
+ "loss": 1.1155,
879
+ "step": 10300
880
+ },
881
+ {
882
+ "epoch": 2.26,
883
+ "learning_rate": 2.7397260273972603e-05,
884
+ "loss": 1.1163,
885
+ "step": 10400
886
+ },
887
+ {
888
+ "epoch": 2.28,
889
+ "learning_rate": 2.7179821700369644e-05,
890
+ "loss": 1.1055,
891
+ "step": 10500
892
+ },
893
+ {
894
+ "epoch": 2.3,
895
+ "learning_rate": 2.6962383126766688e-05,
896
+ "loss": 1.0981,
897
+ "step": 10600
898
+ },
899
+ {
900
+ "epoch": 2.33,
901
+ "learning_rate": 2.6744944553163732e-05,
902
+ "loss": 1.0893,
903
+ "step": 10700
904
+ },
905
+ {
906
+ "epoch": 2.35,
907
+ "learning_rate": 2.6527505979560773e-05,
908
+ "loss": 1.0868,
909
+ "step": 10800
910
+ },
911
+ {
912
+ "epoch": 2.37,
913
+ "learning_rate": 2.6310067405957817e-05,
914
+ "loss": 1.0793,
915
+ "step": 10900
916
+ },
917
+ {
918
+ "epoch": 2.39,
919
+ "learning_rate": 2.609262883235486e-05,
920
+ "loss": 1.0699,
921
+ "step": 11000
922
+ },
923
+ {
924
+ "epoch": 2.41,
925
+ "learning_rate": 2.58751902587519e-05,
926
+ "loss": 1.0628,
927
+ "step": 11100
928
+ },
929
+ {
930
+ "epoch": 2.44,
931
+ "learning_rate": 2.5657751685148945e-05,
932
+ "loss": 1.0506,
933
+ "step": 11200
934
+ },
935
+ {
936
+ "epoch": 2.46,
937
+ "learning_rate": 2.544031311154599e-05,
938
+ "loss": 1.0437,
939
+ "step": 11300
940
+ },
941
+ {
942
+ "epoch": 2.48,
943
+ "learning_rate": 2.522287453794303e-05,
944
+ "loss": 1.0397,
945
+ "step": 11400
946
+ },
947
+ {
948
+ "epoch": 2.5,
949
+ "learning_rate": 2.5005435964340073e-05,
950
+ "loss": 1.0436,
951
+ "step": 11500
952
+ },
953
+ {
954
+ "epoch": 2.52,
955
+ "learning_rate": 2.4787997390737117e-05,
956
+ "loss": 1.0313,
957
+ "step": 11600
958
+ },
959
+ {
960
+ "epoch": 2.54,
961
+ "learning_rate": 2.4570558817134158e-05,
962
+ "loss": 1.0209,
963
+ "step": 11700
964
+ },
965
+ {
966
+ "epoch": 2.57,
967
+ "learning_rate": 2.4353120243531202e-05,
968
+ "loss": 1.0199,
969
+ "step": 11800
970
+ },
971
+ {
972
+ "epoch": 2.59,
973
+ "learning_rate": 2.4135681669928246e-05,
974
+ "loss": 1.0153,
975
+ "step": 11900
976
+ },
977
+ {
978
+ "epoch": 2.61,
979
+ "learning_rate": 2.3918243096325287e-05,
980
+ "loss": 1.0073,
981
+ "step": 12000
982
+ },
983
+ {
984
+ "epoch": 2.63,
985
+ "learning_rate": 2.370080452272233e-05,
986
+ "loss": 1.0046,
987
+ "step": 12100
988
+ },
989
+ {
990
+ "epoch": 2.65,
991
+ "learning_rate": 2.3483365949119374e-05,
992
+ "loss": 0.9931,
993
+ "step": 12200
994
+ },
995
+ {
996
+ "epoch": 2.67,
997
+ "learning_rate": 2.3265927375516415e-05,
998
+ "loss": 0.9957,
999
+ "step": 12300
1000
+ },
1001
+ {
1002
+ "epoch": 2.7,
1003
+ "learning_rate": 2.304848880191346e-05,
1004
+ "loss": 0.9856,
1005
+ "step": 12400
1006
+ },
1007
+ {
1008
+ "epoch": 2.72,
1009
+ "learning_rate": 2.2831050228310503e-05,
1010
+ "loss": 0.9821,
1011
+ "step": 12500
1012
+ },
1013
+ {
1014
+ "epoch": 2.74,
1015
+ "learning_rate": 2.2613611654707543e-05,
1016
+ "loss": 0.9727,
1017
+ "step": 12600
1018
+ },
1019
+ {
1020
+ "epoch": 2.76,
1021
+ "learning_rate": 2.2396173081104587e-05,
1022
+ "loss": 0.9708,
1023
+ "step": 12700
1024
+ },
1025
+ {
1026
+ "epoch": 2.78,
1027
+ "learning_rate": 2.217873450750163e-05,
1028
+ "loss": 0.9737,
1029
+ "step": 12800
1030
+ },
1031
+ {
1032
+ "epoch": 2.8,
1033
+ "learning_rate": 2.1961295933898672e-05,
1034
+ "loss": 0.969,
1035
+ "step": 12900
1036
+ },
1037
+ {
1038
+ "epoch": 2.83,
1039
+ "learning_rate": 2.1743857360295716e-05,
1040
+ "loss": 0.9544,
1041
+ "step": 13000
1042
+ },
1043
+ {
1044
+ "epoch": 2.85,
1045
+ "learning_rate": 2.152641878669276e-05,
1046
+ "loss": 0.9515,
1047
+ "step": 13100
1048
+ },
1049
+ {
1050
+ "epoch": 2.87,
1051
+ "learning_rate": 2.1311154598825833e-05,
1052
+ "loss": 0.9502,
1053
+ "step": 13200
1054
+ },
1055
+ {
1056
+ "epoch": 2.89,
1057
+ "learning_rate": 2.1093716025222877e-05,
1058
+ "loss": 0.9409,
1059
+ "step": 13300
1060
+ },
1061
+ {
1062
+ "epoch": 2.91,
1063
+ "learning_rate": 2.0876277451619918e-05,
1064
+ "loss": 0.9451,
1065
+ "step": 13400
1066
+ },
1067
+ {
1068
+ "epoch": 2.94,
1069
+ "learning_rate": 2.0658838878016962e-05,
1070
+ "loss": 0.9388,
1071
+ "step": 13500
1072
+ },
1073
+ {
1074
+ "epoch": 2.96,
1075
+ "learning_rate": 2.0441400304414006e-05,
1076
+ "loss": 0.9337,
1077
+ "step": 13600
1078
+ },
1079
+ {
1080
+ "epoch": 2.98,
1081
+ "learning_rate": 2.0223961730811046e-05,
1082
+ "loss": 0.9343,
1083
+ "step": 13700
1084
+ },
1085
  {
1086
  "epoch": 3.0,
1087
+ "eval_cer": 5.5303,
1088
+ "eval_gen_len": 13.2935,
1089
+ "eval_loss": 0.776914656162262,
1090
+ "eval_rouge1": 0.0,
1091
+ "eval_rouge2": 0.0,
1092
+ "eval_rougeL": 0.0,
1093
+ "eval_rougeLsum": 0.0,
1094
+ "eval_runtime": 28.2855,
1095
+ "eval_samples_per_second": 70.708,
1096
+ "eval_steps_per_second": 0.283,
1097
+ "step": 13797
1098
+ },
1099
+ {
1100
+ "epoch": 3.0,
1101
+ "learning_rate": 2.000652315720809e-05,
1102
+ "loss": 0.922,
1103
+ "step": 13800
1104
+ },
1105
+ {
1106
+ "epoch": 3.02,
1107
+ "learning_rate": 1.9789084583605134e-05,
1108
+ "loss": 0.8943,
1109
+ "step": 13900
1110
+ },
1111
+ {
1112
+ "epoch": 3.04,
1113
+ "learning_rate": 1.9571646010002175e-05,
1114
+ "loss": 0.8884,
1115
+ "step": 14000
1116
+ },
1117
+ {
1118
+ "epoch": 3.07,
1119
+ "learning_rate": 1.935420743639922e-05,
1120
+ "loss": 0.8858,
1121
+ "step": 14100
1122
+ },
1123
+ {
1124
+ "epoch": 3.09,
1125
+ "learning_rate": 1.9136768862796263e-05,
1126
+ "loss": 0.8831,
1127
+ "step": 14200
1128
+ },
1129
+ {
1130
+ "epoch": 3.11,
1131
+ "learning_rate": 1.8919330289193303e-05,
1132
+ "loss": 0.8857,
1133
+ "step": 14300
1134
+ },
1135
+ {
1136
+ "epoch": 3.13,
1137
+ "learning_rate": 1.8701891715590347e-05,
1138
+ "loss": 0.8825,
1139
+ "step": 14400
1140
+ },
1141
+ {
1142
+ "epoch": 3.15,
1143
+ "learning_rate": 1.848445314198739e-05,
1144
+ "loss": 0.8758,
1145
+ "step": 14500
1146
+ },
1147
+ {
1148
+ "epoch": 3.17,
1149
+ "learning_rate": 1.8267014568384432e-05,
1150
+ "loss": 0.8742,
1151
+ "step": 14600
1152
+ },
1153
+ {
1154
+ "epoch": 3.2,
1155
+ "learning_rate": 1.8049575994781476e-05,
1156
+ "loss": 0.8762,
1157
+ "step": 14700
1158
+ },
1159
+ {
1160
+ "epoch": 3.22,
1161
+ "learning_rate": 1.783213742117852e-05,
1162
+ "loss": 0.8654,
1163
+ "step": 14800
1164
+ },
1165
+ {
1166
+ "epoch": 3.24,
1167
+ "learning_rate": 1.761469884757556e-05,
1168
+ "loss": 0.8697,
1169
+ "step": 14900
1170
+ },
1171
+ {
1172
+ "epoch": 3.26,
1173
+ "learning_rate": 1.7397260273972604e-05,
1174
+ "loss": 0.8648,
1175
+ "step": 15000
1176
+ },
1177
+ {
1178
+ "epoch": 3.28,
1179
+ "learning_rate": 1.717982170036965e-05,
1180
+ "loss": 0.859,
1181
+ "step": 15100
1182
+ },
1183
+ {
1184
+ "epoch": 3.31,
1185
+ "learning_rate": 1.696238312676669e-05,
1186
+ "loss": 0.8616,
1187
+ "step": 15200
1188
+ },
1189
+ {
1190
+ "epoch": 3.33,
1191
+ "learning_rate": 1.6744944553163733e-05,
1192
+ "loss": 0.8519,
1193
+ "step": 15300
1194
+ },
1195
+ {
1196
+ "epoch": 3.35,
1197
+ "learning_rate": 1.6527505979560777e-05,
1198
+ "loss": 0.8493,
1199
+ "step": 15400
1200
+ },
1201
+ {
1202
+ "epoch": 3.37,
1203
+ "learning_rate": 1.6310067405957817e-05,
1204
+ "loss": 0.8478,
1205
+ "step": 15500
1206
+ },
1207
+ {
1208
+ "epoch": 3.39,
1209
+ "learning_rate": 1.609262883235486e-05,
1210
+ "loss": 0.8514,
1211
+ "step": 15600
1212
+ },
1213
+ {
1214
+ "epoch": 3.41,
1215
+ "learning_rate": 1.5875190258751905e-05,
1216
+ "loss": 0.8435,
1217
+ "step": 15700
1218
+ },
1219
+ {
1220
+ "epoch": 3.44,
1221
+ "learning_rate": 1.5657751685148946e-05,
1222
+ "loss": 0.8436,
1223
+ "step": 15800
1224
+ },
1225
+ {
1226
+ "epoch": 3.46,
1227
+ "learning_rate": 1.544031311154599e-05,
1228
+ "loss": 0.8393,
1229
+ "step": 15900
1230
+ },
1231
+ {
1232
+ "epoch": 3.48,
1233
+ "learning_rate": 1.5222874537943032e-05,
1234
+ "loss": 0.8355,
1235
+ "step": 16000
1236
+ },
1237
+ {
1238
+ "epoch": 3.5,
1239
+ "learning_rate": 1.5005435964340076e-05,
1240
+ "loss": 0.8364,
1241
+ "step": 16100
1242
+ },
1243
+ {
1244
+ "epoch": 3.52,
1245
+ "learning_rate": 1.4787997390737118e-05,
1246
+ "loss": 0.8331,
1247
+ "step": 16200
1248
+ },
1249
+ {
1250
+ "epoch": 3.54,
1251
+ "learning_rate": 1.457055881713416e-05,
1252
+ "loss": 0.829,
1253
+ "step": 16300
1254
+ },
1255
+ {
1256
+ "epoch": 3.57,
1257
+ "learning_rate": 1.4353120243531205e-05,
1258
+ "loss": 0.8273,
1259
+ "step": 16400
1260
+ },
1261
+ {
1262
+ "epoch": 3.59,
1263
+ "learning_rate": 1.4135681669928247e-05,
1264
+ "loss": 0.8261,
1265
+ "step": 16500
1266
+ },
1267
+ {
1268
+ "epoch": 3.61,
1269
+ "learning_rate": 1.3918243096325289e-05,
1270
+ "loss": 0.8252,
1271
+ "step": 16600
1272
+ },
1273
+ {
1274
+ "epoch": 3.63,
1275
+ "learning_rate": 1.3700804522722333e-05,
1276
+ "loss": 0.8269,
1277
+ "step": 16700
1278
+ },
1279
+ {
1280
+ "epoch": 3.65,
1281
+ "learning_rate": 1.3483365949119375e-05,
1282
+ "loss": 0.8206,
1283
+ "step": 16800
1284
+ },
1285
+ {
1286
+ "epoch": 3.67,
1287
+ "learning_rate": 1.3265927375516418e-05,
1288
+ "loss": 0.8176,
1289
+ "step": 16900
1290
+ },
1291
+ {
1292
+ "epoch": 3.7,
1293
+ "learning_rate": 1.3048488801913462e-05,
1294
+ "loss": 0.8119,
1295
+ "step": 17000
1296
+ },
1297
+ {
1298
+ "epoch": 3.72,
1299
+ "learning_rate": 1.2831050228310504e-05,
1300
+ "loss": 0.813,
1301
+ "step": 17100
1302
+ },
1303
+ {
1304
+ "epoch": 3.74,
1305
+ "learning_rate": 1.2615786040443576e-05,
1306
+ "loss": 0.8097,
1307
+ "step": 17200
1308
+ },
1309
+ {
1310
+ "epoch": 3.76,
1311
+ "learning_rate": 1.2398347466840618e-05,
1312
+ "loss": 0.8073,
1313
+ "step": 17300
1314
+ },
1315
+ {
1316
+ "epoch": 3.78,
1317
+ "learning_rate": 1.218090889323766e-05,
1318
+ "loss": 0.8016,
1319
+ "step": 17400
1320
+ },
1321
+ {
1322
+ "epoch": 3.81,
1323
+ "learning_rate": 1.1963470319634704e-05,
1324
+ "loss": 0.8068,
1325
+ "step": 17500
1326
+ },
1327
+ {
1328
+ "epoch": 3.83,
1329
+ "learning_rate": 1.1746031746031746e-05,
1330
+ "loss": 0.8034,
1331
+ "step": 17600
1332
+ },
1333
+ {
1334
+ "epoch": 3.85,
1335
+ "learning_rate": 1.1528593172428789e-05,
1336
+ "loss": 0.8051,
1337
+ "step": 17700
1338
+ },
1339
+ {
1340
+ "epoch": 3.87,
1341
+ "learning_rate": 1.1311154598825833e-05,
1342
+ "loss": 0.8043,
1343
+ "step": 17800
1344
+ },
1345
+ {
1346
+ "epoch": 3.89,
1347
+ "learning_rate": 1.1095890410958904e-05,
1348
+ "loss": 0.801,
1349
+ "step": 17900
1350
+ },
1351
+ {
1352
+ "epoch": 3.91,
1353
+ "learning_rate": 1.0878451837355948e-05,
1354
+ "loss": 0.7976,
1355
+ "step": 18000
1356
+ },
1357
+ {
1358
+ "epoch": 3.94,
1359
+ "learning_rate": 1.066101326375299e-05,
1360
+ "loss": 0.7968,
1361
+ "step": 18100
1362
+ },
1363
+ {
1364
+ "epoch": 3.96,
1365
+ "learning_rate": 1.0443574690150033e-05,
1366
+ "loss": 0.7983,
1367
+ "step": 18200
1368
+ },
1369
+ {
1370
+ "epoch": 3.98,
1371
+ "learning_rate": 1.0226136116547077e-05,
1372
+ "loss": 0.7915,
1373
+ "step": 18300
1374
+ },
1375
+ {
1376
+ "epoch": 4.0,
1377
+ "eval_cer": 4.8928,
1378
+ "eval_gen_len": 13.209,
1379
+ "eval_loss": 0.666305661201477,
1380
+ "eval_rouge1": 0.0,
1381
+ "eval_rouge2": 0.0,
1382
+ "eval_rougeL": 0.0,
1383
+ "eval_rougeLsum": 0.0,
1384
+ "eval_runtime": 28.2589,
1385
+ "eval_samples_per_second": 70.774,
1386
+ "eval_steps_per_second": 0.283,
1387
+ "step": 18396
1388
+ },
1389
+ {
1390
+ "epoch": 4.0,
1391
+ "learning_rate": 1.000869754294412e-05,
1392
+ "loss": 0.7887,
1393
+ "step": 18400
1394
+ },
1395
+ {
1396
+ "epoch": 4.02,
1397
+ "learning_rate": 9.791258969341161e-06,
1398
+ "loss": 0.7713,
1399
+ "step": 18500
1400
+ },
1401
+ {
1402
+ "epoch": 4.04,
1403
+ "learning_rate": 9.573820395738205e-06,
1404
+ "loss": 0.7674,
1405
+ "step": 18600
1406
+ },
1407
+ {
1408
+ "epoch": 4.07,
1409
+ "learning_rate": 9.356381822135248e-06,
1410
+ "loss": 0.7697,
1411
+ "step": 18700
1412
+ },
1413
+ {
1414
+ "epoch": 4.09,
1415
+ "learning_rate": 9.13894324853229e-06,
1416
+ "loss": 0.7688,
1417
+ "step": 18800
1418
+ },
1419
+ {
1420
+ "epoch": 4.11,
1421
+ "learning_rate": 8.921504674929334e-06,
1422
+ "loss": 0.7681,
1423
+ "step": 18900
1424
+ },
1425
+ {
1426
+ "epoch": 4.13,
1427
+ "learning_rate": 8.704066101326376e-06,
1428
+ "loss": 0.7649,
1429
+ "step": 19000
1430
+ },
1431
+ {
1432
+ "epoch": 4.15,
1433
+ "learning_rate": 8.486627527723418e-06,
1434
+ "loss": 0.7664,
1435
+ "step": 19100
1436
+ },
1437
+ {
1438
+ "epoch": 4.17,
1439
+ "learning_rate": 8.269188954120462e-06,
1440
+ "loss": 0.7619,
1441
+ "step": 19200
1442
+ },
1443
+ {
1444
+ "epoch": 4.2,
1445
+ "learning_rate": 8.051750380517505e-06,
1446
+ "loss": 0.7658,
1447
+ "step": 19300
1448
+ },
1449
+ {
1450
+ "epoch": 4.22,
1451
+ "learning_rate": 7.834311806914547e-06,
1452
+ "loss": 0.7574,
1453
+ "step": 19400
1454
+ },
1455
+ {
1456
+ "epoch": 4.24,
1457
+ "learning_rate": 7.61687323331159e-06,
1458
+ "loss": 0.7614,
1459
+ "step": 19500
1460
+ },
1461
+ {
1462
+ "epoch": 4.26,
1463
+ "learning_rate": 7.399434659708633e-06,
1464
+ "loss": 0.7564,
1465
+ "step": 19600
1466
+ },
1467
+ {
1468
+ "epoch": 4.28,
1469
+ "learning_rate": 7.181996086105676e-06,
1470
+ "loss": 0.7611,
1471
+ "step": 19700
1472
+ },
1473
+ {
1474
+ "epoch": 4.31,
1475
+ "learning_rate": 6.9645575125027185e-06,
1476
+ "loss": 0.7582,
1477
+ "step": 19800
1478
+ },
1479
+ {
1480
+ "epoch": 4.33,
1481
+ "learning_rate": 6.747118938899762e-06,
1482
+ "loss": 0.7558,
1483
+ "step": 19900
1484
+ },
1485
+ {
1486
+ "epoch": 4.35,
1487
+ "learning_rate": 6.529680365296805e-06,
1488
+ "loss": 0.76,
1489
+ "step": 20000
1490
+ },
1491
+ {
1492
+ "epoch": 4.37,
1493
+ "learning_rate": 6.312241791693847e-06,
1494
+ "loss": 0.7605,
1495
+ "step": 20100
1496
+ },
1497
+ {
1498
+ "epoch": 4.39,
1499
+ "learning_rate": 6.094803218090889e-06,
1500
+ "loss": 0.7516,
1501
+ "step": 20200
1502
+ },
1503
+ {
1504
+ "epoch": 4.41,
1505
+ "learning_rate": 5.877364644487932e-06,
1506
+ "loss": 0.7506,
1507
+ "step": 20300
1508
+ },
1509
+ {
1510
+ "epoch": 4.44,
1511
+ "learning_rate": 5.6599260708849755e-06,
1512
+ "loss": 0.747,
1513
+ "step": 20400
1514
+ },
1515
+ {
1516
+ "epoch": 4.46,
1517
+ "learning_rate": 5.442487497282018e-06,
1518
+ "loss": 0.7512,
1519
+ "step": 20500
1520
+ },
1521
+ {
1522
+ "epoch": 4.48,
1523
+ "learning_rate": 5.225048923679061e-06,
1524
+ "loss": 0.751,
1525
+ "step": 20600
1526
+ },
1527
+ {
1528
+ "epoch": 4.5,
1529
+ "learning_rate": 5.007610350076104e-06,
1530
+ "loss": 0.7515,
1531
+ "step": 20700
1532
+ },
1533
+ {
1534
+ "epoch": 4.52,
1535
+ "learning_rate": 4.790171776473146e-06,
1536
+ "loss": 0.7535,
1537
+ "step": 20800
1538
+ },
1539
+ {
1540
+ "epoch": 4.54,
1541
+ "learning_rate": 4.572733202870189e-06,
1542
+ "loss": 0.7511,
1543
+ "step": 20900
1544
+ },
1545
+ {
1546
+ "epoch": 4.57,
1547
+ "learning_rate": 4.3552946292672325e-06,
1548
+ "loss": 0.7526,
1549
+ "step": 21000
1550
+ },
1551
+ {
1552
+ "epoch": 4.59,
1553
+ "learning_rate": 4.137856055664275e-06,
1554
+ "loss": 0.7416,
1555
+ "step": 21100
1556
+ },
1557
+ {
1558
+ "epoch": 4.61,
1559
+ "learning_rate": 3.920417482061318e-06,
1560
+ "loss": 0.7506,
1561
+ "step": 21200
1562
+ },
1563
+ {
1564
+ "epoch": 4.63,
1565
+ "learning_rate": 3.7029789084583605e-06,
1566
+ "loss": 0.7478,
1567
+ "step": 21300
1568
+ },
1569
+ {
1570
+ "epoch": 4.65,
1571
+ "learning_rate": 3.4855403348554037e-06,
1572
+ "loss": 0.7454,
1573
+ "step": 21400
1574
+ },
1575
+ {
1576
+ "epoch": 4.67,
1577
+ "learning_rate": 3.2681017612524463e-06,
1578
+ "loss": 0.7512,
1579
+ "step": 21500
1580
+ },
1581
+ {
1582
+ "epoch": 4.7,
1583
+ "learning_rate": 3.050663187649489e-06,
1584
+ "loss": 0.747,
1585
+ "step": 21600
1586
+ },
1587
+ {
1588
+ "epoch": 4.72,
1589
+ "learning_rate": 2.833224614046532e-06,
1590
+ "loss": 0.7423,
1591
+ "step": 21700
1592
+ },
1593
+ {
1594
+ "epoch": 4.74,
1595
+ "learning_rate": 2.615786040443575e-06,
1596
+ "loss": 0.7408,
1597
+ "step": 21800
1598
+ },
1599
+ {
1600
+ "epoch": 4.76,
1601
+ "learning_rate": 2.3983474668406175e-06,
1602
+ "loss": 0.7457,
1603
+ "step": 21900
1604
+ },
1605
+ {
1606
+ "epoch": 4.78,
1607
+ "learning_rate": 2.1809088932376606e-06,
1608
+ "loss": 0.7447,
1609
+ "step": 22000
1610
+ },
1611
+ {
1612
+ "epoch": 4.81,
1613
+ "learning_rate": 1.9634703196347033e-06,
1614
+ "loss": 0.7346,
1615
+ "step": 22100
1616
+ },
1617
+ {
1618
+ "epoch": 4.83,
1619
+ "learning_rate": 1.7460317460317462e-06,
1620
+ "loss": 0.7404,
1621
+ "step": 22200
1622
+ },
1623
+ {
1624
+ "epoch": 4.85,
1625
+ "learning_rate": 1.528593172428789e-06,
1626
+ "loss": 0.7494,
1627
+ "step": 22300
1628
+ },
1629
+ {
1630
+ "epoch": 4.87,
1631
+ "learning_rate": 1.3111545988258318e-06,
1632
+ "loss": 0.7361,
1633
+ "step": 22400
1634
+ },
1635
+ {
1636
+ "epoch": 4.89,
1637
+ "learning_rate": 1.0937160252228747e-06,
1638
+ "loss": 0.7417,
1639
+ "step": 22500
1640
+ },
1641
+ {
1642
+ "epoch": 4.91,
1643
+ "learning_rate": 8.762774516199173e-07,
1644
+ "loss": 0.7437,
1645
+ "step": 22600
1646
+ },
1647
+ {
1648
+ "epoch": 4.94,
1649
+ "learning_rate": 6.588388780169603e-07,
1650
+ "loss": 0.7395,
1651
+ "step": 22700
1652
+ },
1653
+ {
1654
+ "epoch": 4.96,
1655
+ "learning_rate": 4.414003044140031e-07,
1656
+ "loss": 0.7393,
1657
+ "step": 22800
1658
+ },
1659
+ {
1660
+ "epoch": 4.98,
1661
+ "learning_rate": 2.2396173081104588e-07,
1662
+ "loss": 0.7436,
1663
+ "step": 22900
1664
+ },
1665
+ {
1666
+ "epoch": 5.0,
1667
+ "eval_cer": 4.7349,
1668
+ "eval_gen_len": 13.2035,
1669
+ "eval_loss": 0.6352977752685547,
1670
+ "eval_rouge1": 0.0,
1671
+ "eval_rouge2": 0.0,
1672
+ "eval_rougeL": 0.0,
1673
+ "eval_rougeLsum": 0.0,
1674
+ "eval_runtime": 27.4711,
1675
+ "eval_samples_per_second": 72.804,
1676
+ "eval_steps_per_second": 0.291,
1677
+ "step": 22995
1678
+ },
1679
+ {
1680
+ "epoch": 5.0,
1681
+ "step": 22995,
1682
+ "total_flos": 1.777894335479808e+17,
1683
+ "train_loss": 0.564045099207411,
1684
+ "train_runtime": 5933.1558,
1685
+ "train_samples_per_second": 1239.969,
1686
+ "train_steps_per_second": 3.876
1687
  }
1688
  ],
1689
  "logging_steps": 100,
1690
+ "max_steps": 22995,
1691
+ "num_train_epochs": 5,
1692
  "save_steps": 500,
1693
+ "total_flos": 1.777894335479808e+17,
1694
  "trial_name": null,
1695
  "trial_params": null
1696
  }